I don't mean "infeasible" or "impractical", I mean it is physically impossible. To get a 50% gain solely through technology improvements we'd have to revoke the laws of thermodynamics and figure out how to change the universal electrical constant. I don't expect to see that happen in my lifetime.I must be part of some elite, because I know that such things are not only happening in my lifetime, they are happening today. As in, right now.
Take electric generation. Over a span of about 40 years from the early 1950's into the 1990's, the standard for electric generation was the coal-fired steam-cycle powerplant. Using perhaps 2 reheats and feedwater heaters, such plants averaged about 33% efficiency. All of this has changed rather suddenly. Most newly-added capacity has been in the form of combined-cycle plants burning natural gas and achieving efficiencies of typically 50% or a bit higher; there is Steven's 50% improvement right there. Gas supplies are becoming tight, but coal-fired technology has not stood still either. One old plant in Terre Haute was repowered with a coal gasifier and gas turbine ahead of the original steam turbine. Not only did this increase its net power output from 90 MW to a whopping 262 MW, it boosted the thermal efficiency to almost 40%. This is only 20% over the original rather than the 50% Steven says is impossible, but it's a hint.
There are other projects on the drawing board, to use gasified coal to operate fuel cells rather than turbines. I found this reference to a Japanese project to do just that (see section 2.4). The projected efficiency of this was expected to be between 50% and 52% in this 1999 paper; I do not have time at the moment to hunt down later reports to see if this was realistic.
I admit that replacement of steam powerplants with MCFCs is a long-term issue, but repowering of aging steam plants might be desirable in the interim; the ability to clean both ash and sulfur out of the fuel gas and consequent near-elimination of particulates and SOx is a good reason to pursue such upgrades regardless of the efficiency gains.
(There are other, less-obvious ways to slash overall fuel consumption by smarter use of what we're already using. Take all of the fuel-ethanol operations across the corn-growing parts of the country. Regardless of how much of a boondoggle it is for the taxpayer, the crime is compounded by the use of natural gas or even propane to fire the stills which turn the fermented mash into high-proof alcohol. How much more sense does it make to force these plants to relocate near existing powerplants and buy low-pressure steam from them to heat their stills? By my calculations, the slight reduction in efficiency of the steam turbine would be offset by many times the reduction in fuel required to heat the stills, and the total fuel required to produce the ethanol would be reduced by roughly half. The environment would be improved further by the use of the off-gas produced in the fermentation step as diluent gas in the powerplant, reducing NOx emissions while burning off the organic compounds which make current ethanol operations such undesirable neighbors.)
Vehicles are another area where improvement is easy. The entire SUV phenomenon can be considered a crime against efficiency, and the drag of the typical tractor/trailer combo could probably be cut in half with proper integration and aerodynamic tweaking; that's not 50% improvement, it's 100%. Moving cargo from trucks to rail would increase efficiency roughly 900%. For smaller vehicles, using diesel powerplants can improve efficiency by 50% or more with minimal changes in technology (my current experiments with a sample size of 1 are showing improvements of 50% on the low end, 100% on the high end).
You don't have to improve your production efficiency if you can reduce what you need. There are still lots of people out there using incandescent lamps. Changing to compact fluorescent reduces power consumption by roughly 75%; this is a 300% improvement which is added on top of any improvements in generation. Even in commercial buildings, incandescent lighting accounts for roughly 25% of all electricity used for lighting. Incandescent bulbs have a typical lifetime of 800-1000 hours, so most of those bulbs could potentially be replaced within 2 years at the most.
If that's not enough you can start getting fancy. If you start replacing gas-fired heating plants with cogenerators things start to get really interesting. The average dwelling which heats with natural gas uses ~50 million BTU per heating season. If you burn this gas in a cogenerator at 25% generation efficiency and the same 95% overall efficiency of typical condensing furnaces, instead of generating 47.5 million BTU of heat you'll make 35 million BTU of heat and 12.5 million BTU of electricity (3660 KWH). Feeding 1220 KWH to a heat pump with a CoP of 3.0 will replace the lost heat and leave 2440 KWH for other purposes. If that same electricity would otherwise have been generated in a coal-fired plant with a heat rate of 10,200 BTU/KWH, the cogenerator would eliminate 25 million BTU of coal consumption with zero increase in gas consumption. That's 50% again, and if you could get more than 25% efficiency out of the cogenerator the results would be even more spectacular. (You can get even fancier than that. A heat pump allows the use of outside electricity for space heat when electricity is available and a seamless switchover to gas when it is not. This is a perfect match for wind power during the winter.)
If 100% or even 300% improvement isn't enough, how about infinity? There are plenty of architectural changes which can be made to buildings (both new and existing) which can substantially reduce their energy needs. For instance, why do we have so many buildings lit by electricity during the day when we could just pipe in sunlight using something like Solatubes? Being able to zero out the lighting needs of the top floors of many buidings, plus substantial portions of lower floors near outside walls, would make a good-sized dent in total electrical requirements. So would full replacement of CRT computer monitors with LCD flat panels, and this change is already under way.
These are some of the reasons why I think Steven Den Beste is being overly pessmistic about the possibilities. I think this is a terrible shame, because one of the biggest reasons that things don't improve is that not enough people realize that it can actually be done. (One can easily be pessmistic about what will happen, because we can always muddle along the current course and refuse to change the way we do things; the failure to understand how things can be improved practically guarantees this result.) I hope I've done my bit to show what can be done and how, and put some momentum on the other side of the scale.
Reader mail welcome - please tell me if you want your letter published here.
Most everyone knows the word entropy, but few know what it means. To the information theoretician entropy is one thing, but to the engineer it is the amount of disorder in a store or stream of energy. Entropy can only be created, not destroyed; this means that any process which accepts or generates entropy eventually has to get rid of it. Entropy is carried by heat (energy), so creating entropy means having to reject heat as heat rather than in some other (potentially more useful) form. The higher the entropy, the more energy must be rejected as heat and the less can be converted to work. Once the entropy of a stream of heat has increased sufficiently there is little you can do except dump it; nothing that operates within the laws of nature can get more out of it.
We can arbitrarily define a zero-point to the entropy of a system, but only changes in entropy really mean something. Total entropy can only increase, and one of the major ways it increases is when heat flows between different bodies. In thermodynamic terms S is entropy and H is enthalpy (heat), and
ΔS = ΔH/TabsIn words, the change in entropy is the change in heat energy divided by the absolute temperature at which the change takes place; transferring heat at high temperatures changes entropy less than transferring heat at low temperatures. Note: all temperatures are absolute (referenced to absolute zero). In English units entropy is stated in BTU/Rankine, and in SI units it is given in Joules/Kelvin.
What does this mean in practice? For one thing, it means that any process which takes heat at a high temperature and lets it become heat at a lower temperature increases the entropy. Designs which emphasize high efficiency work to prevent such things. For instance, the purpose of reheats in a steam-cycle powerplant is to allow a higher pressure in the first boiler. This increases the temperature at which the water boils, which in turn decreases the temperature difference between the combustion gases and the water and decreases the amount of entropy produced in the heat transfer. Feedwater heaters work similarly; they take partially-spent steam (at a lower temperature) out of the stream going through the turbines and use it to pre-heat the water going to the boilers. This both cuts the heat rejected at the condenser ("recycling" it) and decreases the entropy produced by decreasing the temperature difference between the water being heated and the heat being supplied to do the job. The smaller the temperature drop during heat flow, the less entropy is produced and the more energy can be converted to useful forms.
(Entropy can be created by other processes, such as by throttling a stream of fluid from a high pressure to a lower pressure. I am not trying to give an exhaustive list here.)
When you look at delta-entropy as ΔS = ΔH/Tabs, a lot of things that look efficient on the surface are revealed as being horribly sub-optimal. Take the gold standard of home heating, the condensing gas furnace. Some condensing furnaces exhaust a stream of humid flue gas that is barely warm, and boast annual fuel utilization efficiencies (AFUEs) of up to 98%. This sounds really great until you realize that the gas flame inside the furnace may be at 2000 Kelvin while the heated air leaving the unit to the house might be at 350 K or less. The entropy of the gases (heated air and flue gas) leaving the unit is almost six times as great as the entropy of the hot combustion gas just leaving the flame zone. Phrased another way, a large fraction of the energy of the gas could be made to do something else useful before it went to heat air. What kind of things might be useful will be outlined in future posts.
All of the high-efficiency energy generation systems work by keeping entropy increases to a minimum. The combined-cycle system generates the heat in an internal combustion engine (gas turbine) operating at very high temperature; the exhaust from the gas turbine is still hot and boils water to operate a steam turbine, taking a second crack at the heat. Fuel cells work on a different principle, beginning by avoiding the entropy increases inherent in combustion itself and using the chemical combination of fuel with oxidizer to maintain a concentration gradient of chemical species which drives the migration of ions through an electrolyte (the concept of Gibbs free energy becomes important here). Some fuel cells (like molten-carbonate fuel cells and solid-oxide fuel cells) run at temperatures high enough that their waste gas can itself drive a gas turbine, whose waste heat can in turn generate steam for a steam turbine. Total efficiency for such a "triple threat" system might top 80%, and it all works by careful management of entropy.
This is a good thing because there is plenty to blog about, from the foolishness of our government to short-term energy strategy to deal with natural-gas shortages to the confluence of biology and quantum physics in the production of solar cells. More later.
Visits since 2006/05/11: